Evidence for the effectiveness of exercise therapy in knee, hip, hand, and ankle osteoarthritis is inconclusive, with an overview of systematic reviews and randomized controlled trials finding the benefits are negligible or short-lasting.

The overview, published February 17, 2026, in RMD Open, examined five reviews of 100 trials of 8631 participants and 28 supplementary trials involving 4360 participants, the majority of which focused on knee and hip osteoarthritis.

In general, exercise therapy was associated with small reductions in pain but not significant increases in physical function when compared to placebo, sham exercise therapy, or attention control. However, the evidence for effects on both pain and function was of very low certainty.

In knee osteoarthritis, exercise therapy showed very small and short-term effects on pain and physical function when compared to no intervention but again with very low certainty evidence.

When exercise therapy was compared to no intervention for hip osteoarthritis, there were little-to-no differences in pain and physical function in the short-term, with moderate certainty evidence. 

And in the case of hand osteoarthritis, exercise therapy was found with moderate certainty to result in small improvements in pain but not in physical function.

‘Comparative Findings Are Clinically Relevant’

“Evidence of varying certainty and mainly concerned with knee osteoarthritis

suggests that exercise may yield outcomes comparable to other conservative non-pharmacological interventions, pharmacotherapy, and arthroscopic surgery,” the authors wrote. “We found low to moderate certainty that exercise is less effective than knee osteotomy and total joint replacement of the knee and hip in the longer term in selected osteoarthritis patients.”

The study’s first author, physiotherapist Tim Schleimer, a PhD candidate at the Bochum University of Applied Sciences, Germany, said the finding that exercise provides small, transient symptom improvements compared with no care was not entirely surprising. “What was more striking was how strongly effect estimates were influenced by trial size,” Schleimer told Medscape Medical News. 

When researchers removed studies with fewer than 100 participants from the analysis, they found the benefits for pain were even smaller. Similarly, the estimates were also reduced when the analysis was limited to longer-running trials.

The study also found when exercise was compared to total knee replacement for treatment of knee osteoarthritis, the weight of evidence supported knee replacement for moderate improvements in pain and physical function in the long term. “Given that exercise is universally promoted as first-line therapy, these comparative findings are clinically relevant,” Schleimer said.

Choice of Trials’ Comparisons and Variety in Outcomes Make Big Difference

Commenting on the paper, sports physiotherapist Joanne Kemp, PhD, of La Trobe University in Melbourne, Australia, said the results of the review were not surprising, given the wide range of effectiveness seen across hundreds of clinical trials of exercise therapy for knee osteoarthritis. 

Part of the reason for that variation in effectiveness was the variation in the choice of comparison in trials, she said. “If you compare exercise to something very similar to exercise, you’ll find no difference, but that doesn’t mean neither works,” Kemp told Medscape Medical News. “They could both work — it’s just that there’s no difference.”

The small sample size in many trials also meant many were underpowered, Kemp said. Recruitment was challenging for exercise clinical trials, and they were expensive to run because they often involved many exercise and physical therapies, Kemp said. She also noted that adherence was a major problem for trials, and many didn’t report it. 

Another issue was the variety in outcomes. “If you look at exercise trials for strengthening, they will show that they improve strength, because that’s what they’re designed to do,” she said. However, quality of life measures were likely to be much more varied and subjective.

Similarly, there was a variety of exercises that could be used for osteoarthritis, and the effect of someone doing weights three times a week was likely to be very different to the effects of a program of yoga or tai chi, Kemp said. 

Kemp also noted that hip and knee osteoarthritis were combined in the paper. “I get really cranky that people always combine hip and knee together, and they’re totally different,” she said.

Schleimer pointed out that there were still several important questions regarding the role of exercise in osteoarthritis, one of which was whether certain patient subgroups were more likely to benefit from particular exercises or exercise types. He also noted that exercise may have other benefits other than on pain and function, such as improvement cardiovascular health and potentially influencing structural disease progression.

“Exercise remains a valuable option for patients who consider small symptom improvements worthwhile and who are willing and able to engage in activity,” he said. “However, clinicians should communicate realistic expectations.”

No funding or conflicts of interest were declared.